The document discusses how data democratization through an insights marketplace is essential for organizations to become truly data-driven. It defines data democratization as making data accessible across business lines through self-service analytics and predictive platforms. An insights marketplace allows internal users and partners to search, access, and subscribe to shared data assets like reports, models, and raw data. This facilitates collaboration, reduces duplication of efforts, and can help organizations monetize their data internally through improved products and efficiency or externally through partnerships. Examples of Transport for London and educational institutions successfully applying these approaches are provided.
Use AI to Build Member Loyalty as Medicare Eligibility Dates Draw Near
To Become a Data-Driven Enterprise, Data Democratization is Essential
1. Cognizant 20-20 Insights
December 2020
To Become a Data-Driven
Enterprise, Data Democratization
is Essential
To optimise enterprise knowledge, organizations need a modern platform that
enables data to be more easily shared, interpreted and capitalized on by internal
decision makers and by business partners across the extended value chain.
Executive Summary
From the time the term ’data is the new oil’ originated
in 2006, interest across enterprises to become data-
driven has skyrocketed. To become a true data-driven
enterprise, organizations must activate the troves of data
on which they sit to glean insights and foresights that drive
innovation, competitive performance and better customer
experience. This process is called data monetization.
Though the term involves money, it does not always have
to be about selling data or insights. There are two ways
enterprises can monetize data:
❙ Internal monetization, which is about unlocking
the value of the data through innovative products,
operational efficiency or better customer experience.
❙ External monetization, which is about making data
or insights available to partners or other external
consumers for a price (as information services
providers D&B, AC Nielsen, Experian and others do),
or making data or insights freely available so that
consumers can use it to build products (government
entities and educational institutions often do this).
2. Cognizant 20-20 Insights
2 / To Become a Data-Driven Enterprise, Data Democratization is Essential
In our view, data monetization will come automatically
through the democratization of data, and
democratizing data requires a modern data platform.
A modern data platform helps move data from silos
to a common platform, which provides the ability
to bring massive quantities of bits and bytes to
power inductive analytics. The platform also reduces
the time required to extract insights from data,
which accelerates decision-making. By bringing
massive amounts of data and breaking data silos,
organizations can establish a single version of truth
and improve the discovery of data assets. This
discovery is critical to provide data democratization.
Data democratization is the process of enabling
access and availability of information across lines
of businesses to drive innovation via self-service
business intelligence (BI) and predictive analytics
platforms or by applying deep-dive data science. It’s
a departure from the traditional process, in which
data is typically owned by a central IT team and the
lines of business — that is, decision makers — must
work through IT to access the data they need.
In this paper, we lay out our view on the ways and
means of creating an insights marketplace that
modernizes and frees data from its shackles and
provides a way forward for organizations seeking
new and vital ways to innovate and profit from
existing data stores. We conclude with examples
of how democratization is generating results (see
Quick Take, page 9).
3. Typical Challenges Inhibiting Collaboration
Data and analytics shared-services organizations face myriad challenges:
Legacy
Architecture
Data
Platform
vs. Insights
Platform
Unreliable
and Multiple
Versions of
Truth
Lacking
Agility
High
Dependency
Non
collaborative
Unorganized
KPIs and
Assets
Redundant data
and analytical
assets across
multiple
systems
Inability to
gauge usage
and ROI of data
and analytical
assets
Lack of
self-service and
transparency
leads to ‘shadow
IT’ teams
Poor
performance of
systems due to
redundancies
and data
proliferation
Inward looking,
bottom up
approach
on data, no
perspective on
maximization/
monetization
of data assets
internally or
externally
Slow turnaround
of requests
due to low
collaboration
and reuse
among business
users, data
scientists
Insights spread
across BI tools,
models and data,
complicating
seamless access.
Users end up
searching for
KPIs
Cognizant 20-20 Insights
3 / To Become a Data-Driven Enterprise, Data Democratization is Essential
Overcoming Data Democratization Challenges
Building applications directly on modern cloud-
based data platforms comes easily for the digitally
native organizations that have emerged during this
millennium. The advantage is clear: a single, unified
data store that is accessible to a wide community of
consumers instantaneously. However, this approach
represents a huge leap for established enterprises
(see Figure 1).
One goal of data modernization is the creation of a
single trusted data platform that can unify existing
silos, making the combined and standardised data
available to a wide community with clear governance
and security controls in place.
With data modernization, organizations create
a cloud-enabled ecosystem that brings together
data from across the enterprise. This helps teams
cross-pollinate data in ways that uncover actionable
insights and, importantly, makes it available for
consumption on a real-time basis.
A modern data platform consists of three main
capabilities:
❙ A responsive data architecture that is extensible
to changing business and market needs —
meaning it can process a range of the four Vs of
data (volume, velocity, variety and veracity).
❙ Intelligent data management, which includes
proper governance mechanisms and metadata
management.This provides the right level of
controls on the data and renders the data trustable.
❙ Delivery at scale, which encompasses automation
and DevOps methods needed to truly deliver at scale.
Additionally,the modern data platform is set up with
intelligent management capabilities that enable demo-
cratization and monetization of data (see Figure 2).
❙ Data platform: This provides a foundation for
defining a new modern data platform or for
extending a legacy environment to the cloud,where
it can ingest data from a variety of systems of record
within the enterprise and at the desired frequency of
batch,real-time,streaming—or through application
programming interfaces (APIs).
❙ External/internal data: This enables the
ingestion of relevant data wherever it may
reside into the platform via any ingestion
method available (through feeds, for example, or
technology-enabled data exchanges).
Figure 1
4. Cognizant 20-20 Insights
4 / To Become a Data-Driven Enterprise, Data Democratization is Essential
Figure 2
Insights Marketplace Needs a Dedicated Platform
❙ Data governance: This enables the following
critical features:
> Data quality, which helps define and monitor
the quality of the data assets.
> Data catalogue, in which all data assets
are made available along with metadata
and lineage. The data assets also provide
information on the quality service level
agreements (SLAs), thus making the data
more trustable. The metadata available in the
data assets will also include currency stamps
(“use by”).
> Master data, which helps create a single source
of truth for shared master data assets. This
reduces data duplication and increases the
quality of data.
> Data security, which helps establish a role-
based access control (RBAC) mechanism to
govern data access, which must be authorised
by the data owner before access is provided to
the user. It also provides the audit of the users
who have/had access to the data, along with
their level of access.
> Data compliance, which helps implement data
compliance requirements based on the type
of data.
❙ Data catalogue: This provides a repository of
the information available, its lineage, and quality
metrics on the data products to enable data
democratization and self-service.
❙ Data products: Data engineering principles must
be applied to create data products that enable
seamless browsing and consumption.
❙ BI assets: Metadata and tags for the BI and Analytics
assets must be defined and built by users; they can
then be re-used across the enterprise.
❙ Machine learning (ML) assets. Metadata and tags
for re-usable ML models and related feature sets
are defined and built by data science communities
across the enterprise for search and subscription.
❙ Insight marketplace: Democratization is enabled
through a marketplace interface with which users
look for the data assets they need, and which
enables subscription through workflows.
The typical data democratization value chain is
depicted in Figure 3. The goal is to consolidate
data into a modern platform on which internal and
external data is made available for business users,
data scientists, partners, and external consumers.
Each data asset or feature made available to
consumers should be in the form of a data product
that can be consumed in a self-service model.
Set up a
dedicated
modern data
platform or
extend existing
platform
Embed and integrate
external sources into data
landscape
Store, cleanse and
prepare data for
insights generation
Leverage analytics and data science by
building KPIs, reports and dashboards
Perform data science and analytics to address use
cases in service quality, forecasting and preventive
maintenance
Enable democratization
through the insights
marketplace
Implement state-of-the-art
access control, security and
governance mechanisms
5. Marketplace
Store Owner
Vendors
Multiple-Vendors
Products
eCommerce
Marketplace
Customers
Creating an Insight Marketplace
Online retailers such as Amazon and eBay have
perfected the model of self-service in the world of
shopping. A typical online marketplace for products
works as shown in Figure 4.
5 / To Become a Data-Driven Enterprise, Data Democratization is Essential
Figure 4
Multi-Vendor Marketplace Structure
Figure 3
Benefits
to Core
Business
Data Producers
Data
Acquisition RAW
DATA
TRANSFORMED,
CLEANSED, STD.
& ENRICHED
DATA
MI & Insights
ML models
Data &
Information
Products
Data
Management,
Transformation &
Operations
Data
Generation
Data Aggregators
Data Governance
Data Persistence & Infrastructure (HW & SW)
Data Consumers
New
Products
& Revenue
Opportunities
Cognizant 20-20 Insights
The Insights Sharing Value Chain
6. Cognizant 20-20 Insights
6 / To Become a Data-Driven Enterprise, Data Democratization is Essential
This model can be applied directly to a data-driven
enterprise to enable self-service of data products, as
illustrated in Figure 5.
An insights marketplace offers a nice way to
democratize the data within an organization in its
journey toward becoming a data-driven enterprise.
An insights marketplace is an interface whereby users
across the enterprise can search for data products,
BI, analytics and ML models that have been produced
across the organization.The insights marketplace
also makes the process of getting access to the data
seamless by automating many data governance
activities, such as data security and provisioning.
Users will be able to request data assets that are not
available in the marketplace, as well.The insights
marketplace can drive the following (see Figure 6)
(next page):
❙ Collaboration across teams and business units.
❙ A reduction of duplicated efforts to ingest data
and create data products.
❙ Secured data democratization.
❙ Internal monetization of data.
Figure 5
The Data-Driven Enterprise
Internal Data Sources
❙ CRM
❙ Finance
❙ HR
❙ ERP
❙ Customer Services
Data
Platform
Business
users
Partners
Data
Scientists
External
consumers
Insights
Marketplace
Data Governance, Security, eCommerce
External Data Sources
❙ Weather
❙ Social Media
❙ Credit Scoring
❙ Etc.
❙ BI/Reports
❙ Raw Data
❙ Features
❙ AI/ML models
7. Cognizant 20-20 Insights
7 / To Become a Data-Driven Enterprise, Data Democratization is Essential
Enterprises can use the data, which they have
collected and curated, to monetize through external
partners and other consumers. Data sharing can be
based on subscription and agreement between the
enterprise and a specific partner. This sharing can be
done through APIs or data snapshots. This method
of data sharing can be used to add an additional
stream of revenue.
Figure 6
What an Insights Marketplace Begets
A marketplace for enterprise users enabling collaboration and reusability of
information management and analytics assets
• Publish Insights – Authors can publish raw data/reports/dashboards for consumption by other users
• Advanced Search – Search results shows suggestions of published raw data, reports, dashboards
• Request – Recommended authors can be reached for placing insights request that are not available
Insights after search
• Notifications – Personalized notifications when insights requests are completed by authors
• Data Driven Culture – Asks users to contribute and rates contributors and their content (top rated)
• Collaboration – Enables users to share and comment on the insights
• Narrative – Insights about the raw data/report/dashboard are automated and done through
Sciences Narrative Sciences
• Ability to download the raw data/report/dashboard
• Subscription feature enabling users to get notified on insights being refreshed
• Productivity
• Subscribed insights are summarized to highlight key messages without need to click down for details
• Chatbot lets users ask questions, returns specific information and charts
• Personalized Experience
• Users see subscribed, recommended, most downloaded, and top-rated insights as well as recent searches
8. Cognizant 20-20 Insights
8 / To Become a Data-Driven Enterprise, Data Democratization is Essential
Benefits of the Insights Marketplace
Democratizing data assets through an open
environment like an insights marketplace enables
collaboration between the suppliers of data assets
and the consumers of those assets, who could be
internal or external to the enterprise (see Figure 7).
Key benefits of the marketplace to internal
consumers through internal monetization include:
❙ Increased collaboration between teams across the
enterprise.
❙ Complete visibility of all data assets available in
the enterprise.
❙ Availability of all types of data assets in the
enterprise (raw data sets, BI and insights, features,
AI/ML products, and more).
❙ Reduction of the duplication of efforts by multiple
teams.
❙ Quicker innovation, and the ability to build data
products using assets built by different teams in
the enterprise.
❙ Personalised experiences, such as the ability
to associate data product use with personas/
roles to permit recommendations for roles; and
usage billing, which allows data products to be
billed according to their value (in the case of data
monetization) and to be monitored for usage in
all sharing modes.
As noted above, in addition to internal users and
teams, an insights marketplace can provide benefits
through external monetization. Among the key
benefits:
❙ A standard framework to share data with partners,
who can then use the data and insights to
improve their own services or products.
❙ The potential to unlock a data asset, creating a
new revenue stream through sharing data with
different partners or external consumers.
❙ Strong adherence to data security/compliance
rules when sharing the data with external
partners/consumers.
The initial step to unlock the value of the data and
utilise the benefits mentioned above is by bringing
a culture change across the organization in driving
innovation through collaboration and self-service
powered by seamless availability of data. The key is to
make information for business purposes as available
as possible, which enables the business benefits to
quickly accrue.
Figure 7
Key Ecosystem Synergies
Data Products Creation
Create data products from internal and external data
sources and register the assets through the catalog
Configure and Record Assets
Configure search criteria, description of the
asset, validity duration
Cost Visibility
Cost information availability and
easy access for intended parties
Personalization of Services
Provides feature to let supplier
personalize response/interactions with customer
for improved and enriched service experience.
Assets Catalog
A catalog of all available assets by category
Search
Search for data assets based
on popularity or other criteria
Privilege Biased
A given user sees only what
they are authorized to access
Fast & Easy
Access Control
Requesting access in some cases
happens without approval while some go
through Org Governance chart for approval
Suppliers
Consumers
9. Cognizant 20-20 Insights
9 / To Become a Data-Driven Enterprise, Data Democratization is Essential
Data monetization in action
One example of cultural change and data monetization is taking place at Transport for
London (TfL). TfL makes its data available to businesses through its Open Data initiative,
and the results have been nothing short of extraordinary. Through mobile apps, developers
in the public domain access the data to build products (such as route planners and traffic
disruption notifications) that not only provide essential services but help their businesses to
build consumer loyalty.
Another example is in the educational sector. The EU’s Open Data initiative is helping
educational institutions understand job demand and skill gaps in the workforce. This data is
used by educational institutions to create courses that can be utilised by students, thereby
closing the skills gap and meeting job demand.
Quick Take
11. Cognizant 20-20 Insights
11 / To Become a Data-Driven Enterprise, Data Democratization is Essential
About the authors
Vinod Kannan
Data Modernization Consultant, AI & Analytics, Cognizant
Vinod Kannan is a Data Engineering, AI and analytics, and enterprise
architecture specialist within Cognizant’s AI & Analytics practice. With more
than 20 years of experience, he has driven several transformational data
modernization programmes for strategic customers across multiple industry
sectors. He has a degree in Engineering. Vinod and can be reached at vinod.
kannan@cognizant.com | www.linkedin.com/in/vinod-kannan-1519562.
Madhusudhanan Padmanabhan
Enterprise Architect, AI & Analytics, Cognizant
Madhusudhanan Padmanabhan is an Enterprise Architect within Cognizant’s AI &
Analytics practice.He has over 22 years of experience architecting and building digital
platforms for enterprises,as well as in-depth experience and expertise in building
distributed applications and data analytics solutions.Madhu has a master’s degree in
computer application and is a Microsoft Certified Solution Architect Expert.He can be
reached at Madhusudhanan.P@cognizant.com | www.linkedin.com/in/pmadhu/.